Automatic Pavement Crack Detection Based on Hierarchical Feature Augmentation

As one of the most important infrastructures, road is the foundation of transportation. The detection of pavement cracks has become an important task. However, due to the complexity and diversity of crack morphology, detecting cracks in the road has become a challenging task. Poor detection effect and weak generalizability are the problems of most current crack detection methods. Inspired by deep learning related technologies, a crack detection network based on feature augmentation is proposed to extract crack areas in pavement images. The network uses parallel dilated convolution branches to capture different scales image information. The hierarchical feature augmentation module is proposed to combine key information from feature maps of different levels. And some side networks are introduced to perform prediction individually at each level. To demonstrate the performance of the proposed crack segmentation network, experiments are carried out on public datasets. Compared with other advanced crack detection methods, the proposed method can effectively predict the crack pixels in pavement image. The detection accuracy and generalization are improved.